The present invention relates to an information processing device, an information processing method, and a computer program product for retrieving desired objects such as text data, image data, etc. from a database containing a plurality of such objects to display the desired objects.
With the rapid spread of computers in recent years, the amount of electronic information to be handled by people has been increased rapidly. In order to manage the electronic information efficiently, a database containing a large number of pieces of information is prepared to retrieve desired electronic information therefrom as necessary.
Regarding the retrieval of text data (i.e., data including character string) , there is a technique of setting an appropriate keyword as a search key to retrieve text data including a word/phrase corresponding to the keyword from the database. Further, Japanese Patent Laid-Open Publication No. 6-259485 discloses a technique to be used in displaying the retrieved text data on a display device. In this technique, a word/phrase corresponding to a keyword used for retrieval is emphatically displayed in a manner distinguished from other words.
When retrieval of the text data is executed by using a keyword as described above, normally, a plurality of text data (a large number of text data, depending on a case) are retrieved as candidate documents. Thus, there are cases in which a user desires to check the candidate documents as to whether or not they are desired ones.
Hitherto, to determine whether a candidate document is a desired document, a user had to check contents is of the entire candidate document. This is a very burdensome task to the user. In such a situation, there is a growing demand for the development of a technique of allowing a user to easily check whether retrieved candidate documents are desired ones.
In the technique described in the Japanese Patent Laid-Open Publication No. 6-259485, only a word/phrase corresponding to the keyword is emphatically displayed. Thus, the same words or phrases are emphatically displayed in the respective candidate documents. Thus, the technique is not convenient enough to be used to check the contents of the candidate documents.
On the other hand, Japanese Patent Laid-open Publication No. 7-114573 discloses an image retrieval device wherein images, which are objects, are stored along with associated search keys (keywords) serving as attribute information, and an image having a search key or keys coincident with a search key or keys entered as a retrieval constraint is retrieved and displayed.
In the image retrieval device, a plurality of retrieved images are displayed in descending order of degrees of similarity to the retrieval constraint. The degree of similarity is obtained by calculating a function of a precedence correlation between the ranking of the search keys serving as the retrieval constraint ranked by an input order thereof and a ranking indicating an occurrence order of the corresponding search keys annexed to each image.
In the image retrieval device, retrieved images are displayed in the descending order of degrees of similarity to the input retrieval constraint of the attribute information of the retrieved image. Thus, an image desired by a user can be displayed at a higher rank.
However, because in the image retrieval device, the degrees of similarity of displayed images are not shown, only one retrieval operation does not enable the user to determine which are target or desired images when there are many candidate images. Thus, a threshold degree of similarity for determining candidate images is increased and a subsequent retrieval is performed using the increased threshold degree of similarity.
However, as described above, because the degrees of similarity of displayed images are not shown, the user does not know what is a proper threshold for efficiently limiting the candidate images. Thus, the user cannot help setting a threshold as he or she likes. If a large number of candidate images each having a high degree of similarity are registered, however, the user may repeat retrieval operations, although it is unnecessary. Thus, the image retrieval device is not user-friendly, and has a problem that the user cannot retrieve images efficiently.